中国药物警戒 ›› 2025, Vol. 22 ›› Issue (7): 742-748.
DOI: 10.19803/j.1672-8629.20250257

• 细胞和基因治疗产品评价研究专栏 • 上一篇    下一篇

基于深度学习卷积神经网络识别CAR-T细胞产品的研究

任禹珂1, 屈哲, 赖梓漩1,2, 张頔1, 赵永田3, 杨艳伟1, 李双星1, 霍桂桃1, 周晓冰1, 林志1,*, 耿兴超4#   

  1. 1中国食品药品检定研究院安全评价研究所,药品监管科学全国重点实验室,北京市重点实验室,细胞及基因治疗药物质量和非临床研究与评价北京市重点实验室,北京 100176;
    2中国药科大学多靶标天然药物全国重点实验室,江苏 南京 210009;
    3美国Indica数字病理实验室,美国 新墨西哥州 87114;
    4中国食品药品检定研究院生物制品检定所,北京 102629
  • 收稿日期:2025-04-27 出版日期:2025-07-15 发布日期:2025-07-17
  • 通讯作者: *林志,女,博士,研究员,药物临床前安全性评价。E-mail: linzhi@nifdc.org.cn #为共同通信作者。
  • 作者简介:任禹珂,女,在读博士,药物临床前安全性评价。△为共同第一作者。
  • 基金资助:
    国家重点研发计划（2024YFA1107302）; 药品监管科学全国重点实验室课题（2023SKLDRS0127）

Pharmacodynamic Evaluation of CAR-T Cell Products Based on Deep Learning Convolutional Neural Network Recognition

REN Yuke1, QU Zhe, LAI Zixuan1,2, ZHANG Di1, ZHAO Yongtian3, YANG Yanwei1, LI Shuangxing1, HUO Guitao1, ZHOU Xiaobing1, LIN Zhi1,*, GENG Xingchao4#   

  1. 1National Institutes for Food and Drug Control, National Center for Safety Evaluation of Drugs, State Key Laboratory of Drug Regulatory Science, Beijing Key Laboratory, Laboratory of Quality Control and Non-clinical Research and Evaluation for Cellular and Gene Therapy Medicinal Products, Beijing 100176, China;
    2State Key Laboratory of Natural Medicines, China Pharmaceutical University, Jiangsu Nanjing 210009, China;
    3Indica Labs, Inc. 8700 Education Pl, Albuquerque, New Mexico USA. NM 87114;
    4National Institutes for Food and Drug Control, Institute for Biological Product Control, Beijing 102629, China
  • Received:2025-04-27 Online:2025-07-15 Published:2025-07-17

PDF (PC)

37

摘要： 目的 通过深度学习（Deep Learning,DL）技术,建立小鼠肝脏淋巴瘤辅助诊断模型,提高病理诊断的准确性和一致性,助力细胞治疗类产品的研发。方法 收集嵌合抗原受体T细胞（Chimeric Antigen Receptor T-Cell,CAR-T）细胞治疗产品的药效和毒理学研究中小鼠肝脏淋巴瘤组织102例和正常小鼠肝脏组织41例,扫描成数字切片后,进行半自动化数据标注。为提高标注的准确性,对所有数据进行组织提取、伪影去除的预处理后,按照8∶1∶1的比例随机分为训练集、验证集和测试集。应用5种不同的卷积神经网络（Convolutional Neural Networks,CNN）识别肝脏淋巴瘤和非淋巴瘤区域,包括FCN、LR-ASPP、DeepLabv3+、U-Net和DenseNet。基于切片图像的肿瘤预测图像进行比较,并采用精确率（Precision）、召回率（Recall）、F1评分（F1-Score）对构建的算法模型进行性能评估。结果 DenseNet、DeepLabv3+和FCN算法的精确率、召回率和F1评分均接近或超过95%。其中,DenseNet算法模型在测试集中的性能评估最佳,其总体精确率为0.989 4,召回率为0.990 6,F1-Score为0.990 0。结论 本研究建立的DenseNet算法模型对于辅助诊断小鼠肝脏淋巴瘤具有良好的应用前景,可有效提高CAR-T细胞治疗产品药效学和毒理学研究中肿瘤发生的评估准确性和一致性。

关键词: 嵌合抗原受体T细胞, 肝脏淋巴瘤, 小鼠, 人工智能, 深度学习, 药效学评价, 毒性病理学

Abstract: Objective To establish an auxiliary diagnostic model for mouse liver lymphoma using deep learning (DL) technology in order to improve the accuracy and consistency of pathological diagnosis and facilitate the research and development of cell therapy products. Methods A total of 102 hepatic lymphoma tissues and 41 normal liver tissues were collected from mice used in chimeric antigen receptor (CAR) T-cell therapeutic and toxicological studies. After the tissues were scanned into digital slides, semi-automated data annotation was performed. To enhance the accuracy of annotation, all data underwent preprocessing (tissue extraction and artifact removal) and was randomly divided into training, validation, and test sets at a ratio of 8∶1∶1. Five convolutional neural networks-FCN, LR-ASPP, DeepLabv3+, U-Net, and DenseNet-were applied to identify tumor and non-tumor regions. The tumor prediction images based on slice images were compared, and the performance of the constructed algorithm models was evaluated using precision, recall, and the F1-score. Results The precision, recall, and F1-score of the DenseNet, DeepLabv3+, and FCN algorithms were all close to or exceeded 95%. Among them, the DenseNet algorithm model performed best in the test set, with an overall precision of 0.989 4, a recall of 0.990 6, and an F1- score of 0.990 0. Conclusion The DenseNet algorithm model established in this study has a good prospect of application in auxiliary diagnosis of mouse liver lymphoma, which can effectively improve the accuracy and consistency of assessment of tumor occurrence in studies on the efficacy and toxicology of CAR-T cell therapy products.

Key words: CAR-T, Hepatic Lymphoma, Mice, Artificial Intelligence, Deep Learning, Pharmacodynamic Evaluation, Toxicological Pathology

中图分类号: